Analysis of AI-Readiness of University Students Using AI-Competency Measurement Framework

Chinoracky, Roman; Stalmasekova, Natalia; Majercakova, Margita; Neumannova, Rebecca

doi:10.3390/educsci16050692

Abstract

Historically, technological progress has driven shifts in the labour market, leading to the disappearance of certain jobs while simultaneously creating new roles fueled by the need to work with emerging technologies. The technological advancements of the early 2020s are inherently linked to Artificial Intelligence (AI) and the rise in chatbots, whose accessibility and ease of use have become paramount for business development. Given this context, the aim of this study is to analyse frameworks describing the AI competencies of students who will constitute the future workforce. Based on an analysis of existing frameworks, a new framework is formulated through synthesis and operationalized into survey items representing AI-related competencies. These survey items are measured by primary research focused on a sample of undergraduate students at a selected faculty and university. The research provides valuable insights for curriculum development policy by highlighting competencies that students perceive as significant versus those they find less important. Building on these findings, the study offers policy recommendations for curriculum designers. The proposed recommendations enable the creation of educational programmes with relevance to the practical needs of the business sector, increasingly impacted by the emergence of AI.

Keywords:

artificial intelligence; ai-competencies framework; higher education; undergraduate students

1. Introduction

The objective of every higher education institution is to provide education that contributes to the development and cultivation of skills and competencies on the basis of which graduates become employable individuals in the labour market. As companies strive to remain competitive, continuous innovation of business processes takes place, resulting in labour demand that is not static but dynamic. This dynamism is particularly reflected in changing requirements placed on higher-education graduates. One of the factors influencing these changes is technological development associated with disruptive technological innovations.

One such innovation in the contemporary technological landscape is artificial intelligence (AI). AI exists in various forms, and its presence is not new. The year 2022 represents a key milestone, as the deployment of the chatbot ChatGPT and its GPT models fundamentally transformed the practical application of AI technologies. This has led to a reshaping of business processes and, consequently, to changes in the set of work-related competencies required to perform a wide range of occupations.

AI-related skills are no longer merely a positive addition to graduates’ curricula vitae. They have become a core requirement in the contemporary labour market. Graduates who report higher levels of AI knowledge and more frequent use of AI tools tend to perceive greater job opportunities and better career alignment in AI-driven fields (Portocarrero Ramos et al., 2025). On the other hand, employers increasingly emphasise that, in addition to formal degrees, graduates require a combination of technical, cognitive, and soft skills in order to perform effectively alongside AI. OECD reports that as AI becomes more widespread, labour market demand is growing not only for specialised AI-related competencies but also for foundational digital and data science skills, complemented by transversal cognitive skills such as problem-solving, teamwork, and communication (OECD, 2023).

It can be argued that AI-related competencies are becoming increasingly important and that their appropriate conceptualization and categorization are essential for policymakers and curriculum designers. A clear understanding of AI competencies enables the adaptation of curricula and educational programmes to better reflect labour market needs.

At present, several frameworks can already be observed that focus on identifying AI-driven competencies expected of higher-education graduates.

UNESCO’s AI Competency Framework for Students aims to get students ready to use AI in a responsible and creative way. It stresses the importance of foundational knowledge, critical thinking, and ethical awareness as key parts of lifelong learning in the AI era. The framework defines AI competence in four interconnected areas: a human-centred mindset, the ethics of AI, AI techniques and applications, and AI system design. These areas are organised into three levels of progression (Understand, Apply, Create), which make up twelve competency blocks. The human-centred and ethics-focused dimensions emphasise values, accountability, civic duty, and ethical reasoning. This shows UNESCO’s view that being aware of ethics and being socially responsible are important parts of being good at AI. AI techniques and applications focus on data literacy, programming concepts, and practical tool use to improve technical skills. AI system design, on the other hand, focuses on higher-order system-level skills like problem scoping, architecture design, and iterative optimisation. These technical aspects are integrated within a human-centred and ethical framework, emphasising that AI competence must be coupled with critical analysis and responsible decision-making. The framework encompasses the ethical, cognitive, technical, and systemic aspects of AI engagement, while the transition from comprehension to application and innovation facilitates a spiral learning process aimed at the co-creation of AI systems and solutions (Miao et al., 2024).

The University of Michigan Ross School of Business’s LEADERS Framework sees AI proficiency as a roadmap for future AI talent. It is based on seven pillars that work together: literacy, enablement, application, development, ethics, and governance, Research and Refinement, and Society. These pillars cover the whole range of AI skills, from knowing the basics of AI concepts and limitations to using AI tools in a smart and effective way at work, to building and improving AI systems over time. The framework stresses the importance of using AI responsibly by focusing on ethics, governance, and regulatory awareness. At the same time, it puts AI competence in the context of its larger effects on society, the economy, and culture. The LEADERS Framework is a role-agnostic model that can be used by people in many different jobs, from operational staff and managers to developers, researchers, and organisational leaders (Jolly et al., 2024).

The Southern Regional Education Board’s (SREB) AI-Ready Workforce Skills framework defines AI readiness in terms of three related competency groups. Success skills are human-centred, transferable skills like communication, teamwork, leadership, problem-solving, adaptability, creativity, and lifelong learning. These skills are important for using AI effectively in AI-augmented work environments. Industry baseline skills encompass the essential academic, domain-specific, ethical, legal, and cybersecurity knowledge necessary for responsible AI engagement across sectors, underscoring that ethical and contextual comprehension is a prerequisite for employability and accountability. Technical skills emphasise practical comprehension and engagement with AI technologies, encompassing data literacy, fundamental programming concepts, essential AI principles, and familiarity with AI-enabled systems. Technical proficiency is considered pertinent to all learners, not exclusively AI specialists. The framework has a four-level learning progression—Awareness, Exploration, Application, and Fluency—that is used in all competency areas. This progression shows that people are becoming more knowledgeable and independent in their use of AI. It is also designed to be flexible so that people can start at different levels and keep learning and getting new skills throughout their lives (SREB, 2025).

In addition to the institutionally focused frameworks previously mentioned, frameworks catalogued in the scientific databases Web of Science and Scopus were examined. The search queries used are presented in Appendix A, Table A1. Inclusion criteria for identified records were explicit mention of AI competency frameworks of higher education students. The query applied to the Web of Science database identified a total of 13 records, of which only one study provided relevant insights. The query applied to the Scopus database yielded 58 documents; however, none of these met the research focus on AI competencies frameworks of higher education students after detailed screening.

In the study A Competency Framework for AI Literacy: Variations by Different Learner Groups and an Implied Learning Pathway, Chee et al. (2024) put forward a multidimensional AI literacy framework that defines AI competence as encompassing not only technical knowledge but also cognitive, ethical, and career-related aspects, specifically applicable to higher education students. The framework consists of eight interconnected competency domains, encompassing AI fundamentals, data and algorithm literacy, AI interaction and application, AI-assisted problem-solving, ethical and responsible AI, critical assessment of AI systems, career-oriented competencies, and continuous learning. In higher education, a lot of emphasis is placed on the ability to read and comprehend data and algorithms, utilise AI for problem-solving, and critically assess AI-generated outputs. This shows that people need to be careful and know a lot about AI systems before they use them. Adding clear career-related skills links AI literacy to getting a job and working in a certain field. The framework itself sees AI competence as a skill that can be learned and used throughout life. This responds to the fast-changing nature of AI technologies by making adaptability and continuous learning core parts of it. It also gives a conceptual basis for making sure that higher education curricula meet the needs of the job market in AI-augmented environments.

Summarisation

Literature addressing AI-competency frameworks reflects differences in institutional context and intended application:

The UNESCO AI Competency Framework for Students focuses on preparing learners for responsible, human-centred, and ethical engagement with AI, balancing foundational technical understanding with critical judgement and civic responsibility within a lifelong learning perspective.
The LEADERS Framework conceptualises AI competence from a strategic and workforce-oriented perspective, emphasising role-agnostic capabilities that enable individuals to use, apply, develop, and govern AI systems in professional and organisational contexts.
The SREB AI-Ready Skills Workforce framework is explicitly oriented towards employability and workforce readiness, integrating human-centred success skills, baseline AI knowledge, and technical interaction with AI technologies to support effective participation in AI-augmented workplaces.
Chee et al.’s (2024) AI literacy competency framework is an empirically based model for higher education that sees AI competence as a multidimensional construct that includes cognitive, technical, ethical, and career-related skills that are important for college students.

While each of these frameworks addresses AI competencies from a distinct point of view, none of them offers a comprehensive, graduate-oriented synthesis that integrates these perspectives into a unified structure tailored to higher education students’ use of AI. From a knowledge-synthesis perspective, the individual frameworks and their respective dimensions can therefore be summarised in a tabular overview (Table 1). Each dimension or pillar, depending on the terminology used within a given framework, is encoded using a unique identifier (column “Code”).

Table 1. Frameworks for AI competency.

Each of the presented frameworks differs in certain respects; however, it can be argued that their core concepts are closely aligned. By synthesising these frameworks, we propose a new framework where we identify five key competency dimensions that are essential for higher education students when working with AI (Table 2). The connection of individual dimensions and pillars of the analysed framework into the proposed framework is implemented using codes from Table 1.

Table 2. Synthesised framework for AI competency.

The research gap within the context of AI competency frameworks lies not only in identifying core AI competency dimensions for higher education students but also in decomposing these dimensions into empirically grounded factors and translating them into measurable constructs that reflect students’ AI-related competencies and readiness for AI-augmented labour market conditions.

In response to this gap, the present study focuses on the following research objectives:

Identification of key factors constituting each synthesised AI competency dimension (A–E) by examining which dimension-specific factors are empirically relevant and demanded in professional and labour-market contexts.
Operationalisation of dimension factors into measurable items suitable for quantitative data collection.
Collection of data aimed at addressing the question of which AI-related competencies are perceived by students as significant and, conversely, which competencies are perceived as underdeveloped and require further improvement.

2. Literature Review

The first research objective is addressed in the form of a systematic literature review (SLR). The methodological requirements of the implemented SLR are listed in Appendix B.

The rationale for the selection of specific factors, as discussed in the identified studies by SLR, is explained in the following subsections. The transformation of these factors, representing AI-related competency based on SLR outputs, into survey items is presented at the end of each subsection devoted to the respective examined dimension. The survey items are used as the empirical basis of the subsequent research presented in the Results of this study.

2.1. Dimension A—Understanding AI (AI Literacy)

Across multiple scientific studies, a recurring theme emerges: AI is becoming a baseline component of both professional and academic student activities. Graduates’ readiness is no longer determined merely by their ability to use a specific tool but by whether they understand what AI is, where its application is meaningful, the principles it operates on, and the risks it entails.

The first thematic area identified across studies is basic conceptual understanding of artificial intelligence (Table 3, column Survey items, row A1). Multiple works demonstrate that mere exposure to AI tools or their use in academic settings does not automatically imply that students understand what AI is and which types of tasks it can address. Research concentrating on generative AI (Constantin & Iacob, 2025; Reich, 2025; Schmidt et al., 2025) consistently emphasises the differentiation between the functional application of AI and the comprehension of its capabilities and constraints. Similarly, more macro-oriented studies (Tolegen et al., 2025; Grama & Todericiu, 2025) identify AI as a universal future competence while implicitly suggesting that its significance lies precisely in graduates’ ability to understand where and why AI is applicable.

Table 3. Survey items for Dimension A—understanding AI (AI literacy).

The second cluster of findings points to applied understanding of AI use in professional practice (Table 3, column Survey items, row A2). Research examining employability and future skills (Tolegen et al., 2025; Opesemowo et al., 2025; Grama & Todericiu, 2025) consistently highlights that AI use is regarded as a cross-sectoral competence, rather than being limited solely to IT professions. At the same time, empirical research indicates that students often struggle to articulate concretely how and where AI is used in corporate practice. Curricular analyses of applied AI in the financial sector (Porjazoska Kujundziski et al., 2024) further demonstrate that the ability to connect theoretical knowledge of AI with specific application scenarios constitutes a key bridge between education and employability.

The third thematic area, which appears mostly implicitly yet consistently in the literature, concerns understanding the basic principles underlying AI and algorithmic learning (Table 3, column Survey items, row A3). Although most studies do not require students to possess deep technical knowledge, several works (El Hosayny et al., 2025; Reich, 2025; Schmidt et al., 2025) warn that insufficient understanding of the distinction between machine learning and rule-based programming leads to overreliance on AI outputs, automation bias, and uncritical use of AI tools.

The fourth thematic area across studies is understanding the social and ethical implications of AI use (Table 3, column Survey items, row A4). Qualitative studies examining the incorporation of AI into higher education environments (Schmidt et al., 2025; Yurdunkulu et al., 2025) highlight concerns regarding academic integrity, the dependability of AI-generated outputs, privacy safeguarding, and accountability for AI-assisted decisions. Simultaneously, studies on employability and the future of work underscore that ethical contemplation regarding AI is increasingly viewed as a fundamental aspect of professional competence rather than merely a moral position.

2.2. Dimension B—Technical and Digital Skills

Using data and analytical tools is a key technical skill that is directly related to getting a job after graduation (Table 4, column survey items, row B1). Research demonstrates that genuine, data-centric learning substantially enhances students’ self-assurance and perceived data-processing abilities (Morgan et al., 2025; Başal et al., 2025). Analyses of the labour market consistently highlight proficiency in spreadsheet and statistical software as a highly sought-after skill, especially in engineering and analytical positions (Fleming et al., 2024). Nonetheless, comparative analyses reveal a continual disparity between students’ self-assessed digital competencies and the actual demands of the labour market, particularly in data analytics (Bhatti et al., 2022; Aljaaidi et al., 2025).

Table 4. Survey items for Dimension B—technical and digital skills.

Experience with data visualisation naturally extends analytical skills and serves as an important connection between technical analysis and result communication (Table 4, column Survey items, row B2). Empirical research demonstrates that converting analytical outputs into visual representations facilitates enhanced data comprehension and interpretation (Roberts, 2025). In employability-focused and predictive modelling research, visualisation serves as an implicit yet vital element of regression and clustering workflows, facilitating pattern interpretation and informed decision-making (Mahdin et al., 2025; José-García et al., 2023). In conclusion, visualisation is an essential component of data-oriented thinking rather than merely an ancillary technical skill.

Understanding programming and algorithmic logic constitutes a significantly transferable skill beyond the realm of computer science (Table 4, column Survey items, row B3). Experimental evidence indicates that text-based programming more effectively facilitates algorithmic thinking and knowledge transfer compared to graphical methods (Vrbančič & Kocijančič, 2024). Labour market studies show that knowing languages (e.g. Python or MATLAB, versions not specified in the study), which are not always needed, can lead to higher pay (Fleming et al., 2024). In the realm of generative AI, even rudimentary AI utilisation necessitates comprehension of fundamental algorithmic processes, with algorithmic thinking deemed essential for adaptation to technologically complex work settings (Zhan & Wang, 2024; Aly et al., 2025).

Practical use of AI tools is among the most consistently identified technical skills in the analysed studies (Table 4, column Survey items, row B4). Research evidence shows widespread student use of generative AI tools for learning, content creation, and analytical tasks (Ravšelj et al., 2025; Aly et al., 2025). Quantitative studies link more frequent and purposeful AI use to improved employability, productivity, and job–field alignment (Portocarrero Ramos et al., 2025; Hrnjić et al., 2025). Research also warns that superficial AI use without conceptual understanding may slow down long-term competence development (Göltl et al., 2024).

Understanding machine learning fundamentals is identified as an advanced yet increasingly important technical competence (Table 4, column Survey items, row B5). Applied research demonstrates that students can utilise data-driven techniques such as regression and clustering beyond their primary fields of study (Mahdin et al., 2025; José-García et al., 2023). To use intelligent systems well, you need to know how they adapt, especially when it comes to personalised feedback and decisions based on data (Zhan & Wang, 2024). Studies centred on employability indicate that familiarity with AI and machine learning principles is becoming increasingly vital for sustained engagement in the increasingly digitalised economy (Toma & Hudea, 2024; Portocarrero Ramos et al., 2025).

2.3. Dimension C—Cognitive and Analytical Abilities

In AI-mediated settings, effective problem-solving increasingly necessitates data-driven analysis, critical evaluation, and validation of AI-generated outputs, rather than mere acceptance (Table 5, column Survey items, row C1). Studies show that automation increases the likelihood of uncritical acceptance of erroneous recommendations, especially in contexts of uncertainty and insufficient data (Dumitru & Halpern, 2023; Rampersad, 2020; He, 2025). Insufficient analytical reasoning further limits students’ capacity to differentiate correlation from causation in AI-generated data, highlighting data-oriented problem-solving as a fundamental AI-related competency (Xu et al., 2022).

Table 5. Survey items for dimension C—cognitive and analytical abilities.

As AI takes over more decision-making tasks, logical and systematic reasoning becomes a way to balance out what algorithms say (Table 5, column Survey items, row C2). Research shows that effective AI-supported decision-making requires critical evaluation of recommendations and evidence-based judgement, as users must assess not only outcomes but also underlying decision processes (Waring, 2024; Dumitru & Halpern, 2023). The absence of systematic evaluation and logical verification is associated with higher decision error rates and uncritical adoption of AI outputs (He, 2025). Studies further identify evidence-based and analytical decision-making as key predictors of adaptability in AI-intensive work environments, despite being among the most demanded yet underdeveloped skills (Rampersad, 2020; Bartolomé et al., 2025; Wong & Abdullah, 2025).

Interpreting AI outputs constitutes a major cognitive challenge, as algorithmic recommendations are typically presented through statistical models, visualisations, and probabilistic information (Table 5, column Survey items, row C3). Research shows that while students perceive generative AI tools as helpful, they often lack the ability to assess the accuracy, reliability, and significance of AI-generated outputs (Ravšelj et al., 2025). Insufficient analytical and numerical literacy increases overreliance on visually persuasive but methodologically flawed results, particularly when critical interpretation of graphs and statistical indicators is limited (Dumitru & Halpern, 2023; He, 2025). Empirical evidence substantiates the necessity of systematically cultivating AI-output interpretation skills to facilitate informed decision-making predicated on algorithmic recommendations (Xu et al., 2022; Bartolomé et al., 2025).

Critical reflection on the social, ethical, and economic implications of AI signifies an advanced cognitive competence that transcends mere technical usage of AI (Table 5, column Survey items, row C4). As AI systems become more autonomous, the potential for negative effects on job markets and social inequalities rises, requiring a comprehensive assessment of benefits and risks for sustained employability (Waring, 2024). Although students recognise risks associated with tools such as ChatGPT, they often lack the analytical capacity for their systematic assessment (Ravšelj et al., 2025). Without deliberately developed critical thinking, students struggle to evaluate bias, ethical issues, and societal consequences, leading to acceptance of flawed algorithmic inferences and weakened human oversight (Velu, 2023; He, 2025). Evidence from technical and vocational education further indicates that critical reflection on digital technologies remains underdeveloped despite high labour market demand (Wong & Abdullah, 2025).

2.4. Dimension D—Soft Skills and Attitudes Towards AI

In various disciplinary and national contexts, soft skills and personal attitudes consistently emerge as pivotal facilitators of effective AI adoption (Table 6, column Survey items, row D1). AI-enhanced environments rely less on technical expertise and more on a willingness to learn, a drive for ongoing improvement, collaboration, and innovative interaction with technology (Tsiligiris & Bowyer, 2021; Rampersad, 2020). Empirical evidence indicates that willingness to learn about AI strongly influences both AI adoption intentions and perceived employability, independent of prior technical expertise (Al Zaabi et al., 2023; Creely et al., 2025). Therefore, openness to lifelong learning is a core component of digital readiness (Ferro et al., 2023). On the other hand, not being ready for AI to learn all the time is seen as a big risk in digital transformation processes (Pronello, 2026; Mahazan & Ismail, 2025).

Table 6. Survey items for dimension D—soft skills and attitudes towards AI.

Communication skills are widely recognised as important for employability but are scarcely examined in relation to AI-specific knowledge transfer (Bedoya-Guerrero et al., 2024; Rampersad, 2020). The reviewed studies do not directly examine individuals’ confidence in communicating or explaining AI-related concepts, as AI-focused research predominantly emphasises attitudes, willingness, or perceived usefulness rather than communicative self-efficacy (Al Zaabi et al., 2023; Creely et al., 2025). Nevertheless, related evidence indicates that self-efficacy mediates the relationship between AI tool utilisation and learner engagement, as the capacity to comprehend and articulate AI concepts augments perceived control over learning (Singh et al., 2026). Studies in educational settings demonstrate that AI-enhanced environments can bolster communicative confidence by facilitating the clarification of complex concepts (Mahazan & Ismail, 2025). Consequently, the ability to explain AI processes, limitations, and outcomes emerges as an important yet under-researched competence in AI-supported and interdisciplinary settings (Table 6, column Survey items, row D2).

Collaborative engagement consistently emerges as a key dimension across the analysed studies (Table 6, column Survey items, row D3). Teamwork is recognised as a fundamental employability skill that fosters innovation, improves learning outcomes, and enhances adaptability in technology-driven and AI-influenced environments (Bedoya-Guerrero et al., 2024; Rampersad, 2020). Digitally mediated and work-integrated learning suggests that collaboration enhances innovative capacities, while pedagogical research underscores teamwork as fundamental to design thinking and maker-based methodologies for fostering problem-solving and creativity (Rampersad, 2020; Velu, 2023; Silva et al., 2020). Additional evidence highlights the increasing significance of cooperative, interdisciplinary collaboration in addressing AI-driven socio-technical challenges, with digital tools enabling shared knowledge creation and collective innovation (Pronello, 2026; Singh et al., 2026).

Motivation to cultivate AI and digital skills serves as a critical intermediary between access to technological resources and measurable employment outcomes (Table 6, column Survey items, row D4). Through studies, individuals engage more actively with AI technologies, sustain learning efforts, and are more likely to convert technological exposure into labour market advantages. Empirical modelling confirms that internal motivation, autonomy, and self-efficacy are pivotal in converting AI-enabled learning into academic achievement and employability, establishing motivation as a primary catalyst for AI-related competence enhancement rather than a subordinate factor (Singh et al., 2026). This corresponds with extensive research on professional development, which suggests lifelong learning orientation and self-directed skill enhancement as crucial for employability amidst rapid technological advancements (Tsiligiris & Bowyer, 2021; Ferro et al., 2023). Conversely, low motivation for upskilling and reskilling is identified as a significant risk factor that may deepen labour market inequalities in AI-driven economies (Pronello, 2026). Systematic reviews in digital education emphasise that the motivation to learn new technologies is as essential as the availability of technological resources (Mahazan & Ismail, 2025).

Creativity emerges as a complementary human capability that strengthens, rather than competes with, effective AI use (Table 6, column Survey items, row D5). Across the analysed literature, creativity is consistently linked to innovation and technology-mediated problem-solving, even when not addressed explicitly in AI-specific terms. Educational research indicates that AI-enhanced environments promote creative exploration and ideation when creativity is regarded as essential to technological efforts (Velu, 2023). Simultaneously, labour market studies emphasise that creative thinking and non-routine problem-solving remain essential human advantages under increasing automation (Rampersad, 2020). Evidence further suggests that digital technologies can amplify creativity and higher-order thinking, while creative and critical thinking mitigate negative adaptation effects of AI automation and support innovative behaviour (Botha & Dreyer, 2022; Pronello, 2026; Singh et al., 2026). Systematic reviews additionally position creativity as a key mediator between AI use and deep learning outcomes (Mahazan & Ismail, 2025).

2.5. Dimension E—Labour Market Readiness

Graduates’ readiness for AI-enabled work environments is shown to be uneven and frequently insufficient, including in technologically oriented fields (Table 7, column Survey items, row E1). Quantitative assessments in Industry 4.0 contexts indicate that students typically reach only moderate or partial readiness, revealing persistent gaps between educational outcomes and labour market expectations (Wong & Abdullah, 2025; Islam, 2022; Bondin & Zammit, 2025). Evidence from engineering, technical, and vocational education and training, and applied science programmes further suggests that although technical knowledge is often present, students lack confidence in performing AI-related tasks in real professional settings (Camacho-Zuñiga et al., 2025; Kamaruzaman et al., 2025). Sector-specific research, particularly in healthcare education, similarly indicates low self-assessed AI readiness despite high perceived relevance, emphasising the necessity of evaluating perceived capability rather than deducing competence solely from curricular-based exposure (Al Zaabi et al., 2023). Employability studies demonstrate that although employers are seeking AI-related skills, these expectations do not uniformly translate into individual task readiness (Corrales-Herrero & Rodríguez-Prado, 2024; Malikaa & Badra, 2025).

Table 7. Survey items for dimension E—labour market readiness.

Adaptability and a commitment to lifelong learning consistently emerge as key factors influencing sustainable employability in the era of AI (Table 7, column Survey items, row E2). AI-driven job markets change tasks and skills quickly, so workers need to keep learning new things (Corrales-Herrero & Rodríguez-Prado, 2024; Islam, 2022). Empirical studies indicate that adaptability-related competencies—such as openness to learning, curiosity, and flexibility—are rated as equally or more important than technical skills in readiness frameworks (Wong & Abdullah, 2025). Yet, evidence shows a consistent difference between technological exposure and adaptive readiness, as even digitally adept students frequently express reduced self-efficacy and confidence in managing AI-driven transformations (Liu et al., 2025; Alam et al., 2025). Qualitative findings further show that flexible, challenge-based, and interdisciplinary learning environments strengthen perceived readiness for labour market transitions, while traditional instructional models tend to reinforce static skill acquisition (Camacho-Zuñiga et al., 2025; Alam et al., 2025).

The analysed studies reveal how effectively higher education supports labour market readiness (Table 7, column Survey items, row E3). Quantitative evidence shows significant differences in job readiness between institutions, even within the same disciplines, highlighting the role of curricular design, pedagogy, and applied learning opportunities beyond individual effort (Wong & Abdullah, 2025). Students consistently associate employability development with practice-oriented formats such as internships, interdisciplinary projects, and collaboration with external partners (Bartolomé et al., 2025). Conversely, skills gap studies reveal that formal education frequently fails to meet labour market demands, necessitating self-directed learning as an essential supplement to institutional offerings (Islam, 2022; Matiwane & Olaitan, 2025).

The capacity to identify, access, and utilise learning and career-support resources is an inadequately examined yet essential element of labour market readiness (Table 7, column Survey items, row E4). Studies indicate that knowledge of career guidance, training opportunities, and information channels significantly affects graduates’ ability to adapt to changes in the labour market (Liu et al., 2025). Research on continuous learning suggests that readiness for the AI era relies not only on current skills but also on the ability to identify and obtain new skills as technologies advance (Camacho-Zuñiga et al., 2025). Limited awareness of training pathways and educational resources is therefore identified as a barrier to effective AI adoption, even when motivation is present (Al Zaabi et al., 2023; Corrales-Herrero & Rodríguez-Prado, 2024).

3. Materials and Methods

Each of the survey items assigned to dimensions A–E (Table 3, Table 4, Table 5, Table 6 and Table 7) can be formulated as a question within a questionnaire survey. The transformation of individual survey items into the questions applied in the questionnaire is detailed in Appendix C. The final version of the questionnaire is in Appendix D. The questionnaire (Appendix D) conducted in this study consisted of two sections, which in sum had 30 questions. The first section (Table A5) was dedicated to sample demographics. The second section (Table A6) consisted of substantive questions. These questions represented the core of this study. Respondents answered these questions by assigning a value of disagreement or agreement with a statement on a 5-point Likert scale, where 1 represented ‘strongly disagree’ and 5 represented ‘strongly agree’.

A total of 252 respondents participated in the research, representing the study sample. This sample consists of first-year students (preparing for professional practice) from the University of Zilina, Faculty of Operation and Economics of Transport and Communications. Targeting this specifically defined sample aimed to capture the perspectives of incoming students in order to ensure the relevance of the findings for curriculum designers and their future decision-making. The total population consists of 669 individuals. Based on sample calculations, which required a minimum of 245 respondents (at a 95% confidence level and 5% margin of error), it can be said that an adequate number of responses (252) were obtained. The 252 responses in the sample meet the requirements for statistical significance and accurately represent the defined population. Before sending out the questionnaire to the respondents, it was tested on a group of 10 individuals. Their feedback helped improve the wording of some questions so that the target group could understand them better. In the context of missing data, it should be noted that only one questionnaire response was excluded from the analysis because the respondent completed only 24% of the questions (Appendix C, questions no. 9–21) from the core section. Other questionnaire responses with only a small number of missing answers were included in the data analysis.

The demographics of the sample (Table 8) indicate that, in terms of gender, 40% of the participants were male and 60% were female. The sample primarily consisted of Slovak students aged 19 and 20 who were primarily enrolled in secondary and grammar schools. A portion of the students were of international origin, primarily from Ukraine and Kazakhstan. Most of the students who took part in this study had significant prior exposure to digital technologies during their previous studies, and more than 60% have high levels of English proficiency (B2 and C1, C2 level). The sample also consists predominantly of students aiming for future employment in professional fields integrated with AI applications.

Table 8. Sample demographics.

The synthesised framework presented in this study consists of survey items examining respondents’ attitudes on a scale. On this basis, it was essential to assess its internal consistency through reliability analysis. For this purpose, Cronbach’s alpha coefficient was utilised to indicate the extent to which items within a single dimension measure the same underlying construct. According to established standards for social science research, a coefficient value above 0.70 was seen as the minimum level of reliability. Along with the overall Cronbach’s alpha, the Corrected Item–Total Correlation (CITC) was calculated for each question to see how well it fit with the underlying idea. Values of correlation greater than 0.30 show that each item makes a major distinction in how similar the dimension is. The measurement results are detailed in Table 9.

Table 9. Cronbach’s alpha and CITC.

The Cronbach’s alpha values for all dimensions exceed the value 0.70, indicating good to very good reliability and confirming the internal stability of the framework. These results demonstrate that the individual survey items within each dimension are interrelated and consistently measure the intended constructs.

In the context of Research Objective 3, as stated at the end of the introduction, statistical inference was employed to evaluate the empirical data. A key prerequisite for selecting appropriate inferential statistical methods is assessing the nature of the data distribution. To verify the normality of the distribution for each of the five dimensions (A–E), the Shapiro–Wilk test was applied (Table 10). Given that the results indicate a non-normal distribution, these measurements are supplemented by descriptive characteristics of the distribution (Table 11). Skewness and kurtosis values within the interval of −1 to 1 were considered sufficiently close to normal for the purposes of conducting the t-test. For Dimension D, the kurtosis value (1.02) slightly exceeds the threshold of 1; however, this minor deviation was taken into account and considered acceptable.

Table 10. Shapiro–Wilk test.

Table 11. Descriptive characteristics of distribution.

Although the Shapiro–Wilk test (Table 10) indicated a non-normal distribution across all dimensions (p < 0.001), analysis of the distribution and skewness values (Table 11) confirmed that the deviations were not extreme. Given the sample size (n = 252) and the research objective no. 3 (to examine competencies perceived as significant by students), a parametric one-sample t-test was employed. This approach is consistent with the Central Limit Theorem, which ensures the stability of sample means in sufficiently large samples, thereby allowing for the reliable identification of key and deficient areas of student competencies.

The rationale for employing a one-sample t-test is that it facilitates the assessment of whether the perceived significance of a specific competency differs statistically from a neutral position. In the context of the applied 5-point Likert scale, 3 represents the theoretical midpoint, indicating an indifferent attitude (neither agree nor disagree). The use of the t-test enables us to scientifically substantiate that the mean score for a given competency is not merely the result of random fluctuation around this midpoint but represents a clearly defined and statistically significant stance held by the students.

For each competency (survey item) and each dimension (within both the overall and survey item-level analysis), a pair of statistical hypotheses was established and tested:

Null Hypothesis: The mean level of perceived significance for the given competency is equal to the theoretical midpoint. This implies that students perceive the competency neutrally, or their level of agreement is due to random chance.
Alternative Hypothesis: The mean level of perceived significance is statistically significantly higher than the theoretical midpoint.

If the calculated p-value is lower than the established significance level (0.05), the null hypothesis is rejected in favour of the alternative hypothesis. In such cases, the competency is considered to be perceived as statistically significant by the students. The lower the p-value (p < 0.001), the higher the confidence that the result is not due to random chance. In instances where the p-value approaches or is lower than 0.05, we identify areas where the students’ attitudes are less pronounced. These areas are classified as less developed, requiring further improvement.

Statistical significance (p-value) alone indicates only that the observed difference between the mean and the neutral midpoint is not random. However, given the established sample size (n = 252), the p-value may reach significance even with minimal and practically negligible deviations. For this reason, we included the calculation of Cohen’s d in the analysis to measure the effect size (practical significance) of the findings.

Within the context of the research objective, the effect size is examined on two levels:

Values of Cohen’s d > 0.80 are interpreted as a high level of perceived importance for a given competency, whereas values within the interval of 0.50 to 0.79 are considered to indicate a moderate level of agreement.
Closer proximity of students’ attitudes to the neutral midpoint is indicated by lower Cohen’s d values (specifically those falling within the 0.20 to 0.49 range). Competencies with a lower effect size are identified as being perceived as underdeveloped or requiring further improvement.

4. Results

Results of statistical significance (p < 0.001) and a large effect size (d > 0.80) are exhibited across all dimensions (Table 12). These results indicate that all examined areas of AI competencies are perceived by students as significant in terms of their own proficiency. In other words, it is perceived by the students that they possess the specific competencies associated with the utilisation of AI.

Table 12. One-sample t-test and Cohen’s d overall results.

While the statistical significance of all five dimensions as an aggregate was confirmed by the preceding analysis, their decomposition into individual survey items is essential for a more comprehensive fulfilment of the research objective. The aggregate score of a dimension may mask the internal variability of the respondents’ attitudes. This ‘atomized’ perspective allows for a more precise specification of areas requiring further improvement, which would otherwise remain hidden if evaluated only at the level of entire dimensions. Consequently, each item is subjected to a separate one-sample t-test and an effect size calculation.

The results of the Dimension A (understanding of AI) analysis indicate that students’ basic AI competencies are perceived as being significantly above average. A statistically significant difference from the neutral midpoint (p < 0.001) is exhibited by all four items, which confirms their subjectively perceived importance and level of development (Table 13).

Table 13. Dimension A: one-sample t-test and Cohen’s d results.

Within Dimension A, the highest perceived significance is associated with competency A4 (ethical and social awareness of AI), which shows a very large effect size (d = 1.25), positioning it as a clear priority aligned with growing societal emphasis on responsible technological development. There are also high effect sizes for competencies related to understanding the difference between machine learning and classical programming (A3, d = 1.10) and using AI in business and the economy (A2, d = 1.00). This shows that students are very confident in both the basic principles of AI and how they can be used in the market. Conversely, the lowest (albeit still moderate) effect size is observed for fundamental AI comprehension (A1, d = 0.74). This pattern shows an interesting contradiction: students say they are more confident in the practical and ethical parts of AI than in its basic ideas. This means that basic AI ideas need to be made clearer and more consistent.

The results for dimension B of the t-test indicate that while all competencies are perceived as statistically significant, the level of their proficiency and perceived importance (as measured by Cohen’s d) is found to be considerably heterogeneous (Table 14).

Table 14. Dimension B one-sample t-test and Cohen’s d results.

Item-level analysis of Dimension B indicates pronounced differences in students’ technical preparedness. The most important skill is B1, which is basic digital data skills and spreadsheet proficiency. This has a large effect size (d = 1.03), which means that the students have a strong technical background. B2, data visualisation and graph creation, on the other hand, has the lowest effect size in the survey (d = 0.42), which means it is the least developed area, even though it is statistically significant (p < 0.001). Moderate deficits are evident in B3, basic programming and algorithmic thinking (d = 0.68), and B4, practical experience with AI tools (d = 0.69), both indicating medium effect sizes. These findings indicate that students exhibit conceptual understanding; however, their technical implementation and practical application of AI are comparatively deficient. In contrast, B5, which is a basic understanding of machine learning principles, has a large effect size (d = 0.86), which suggests a predominantly conceptual rather than practical orientation.

The results for dimension C of the one-sample t-test confirm that a statistically significant positive attitude is exhibited by students across all four competencies (Table 15).

Table 15. Dimension C one-sample t-test and Cohen’s d results.

Analysis of Dimension C shows that the highest perceived significance is associated with C4, critical evaluation of digital technologies and their societal impact, with an effect size close to the large threshold (d = 0.96), indicating this area as a key student strength. Additionally, a high effect size was observed in C3, which is the interpretation of data outputs, graphs, and statistics (d = 0.89). This is different from Dimension B, where problems were found with making data visualisations. This pattern indicates that students are more confident in interpreting finished outputs than in creating them. C1, data-driven analytical problem-solving (d = 0.68), and C2, logical and structured decision-making (d = 0.79), have the lowest effect sizes in this dimension. This means that these are the areas with the greatest room for improvement. Overall, the results show that the students have a strong base in critical thinking and evaluation, but their practical analytical problem-solving skills are still not as strong.

High levels of both statistical and practical significance are exhibited by measurements for Dimension D. t-test results position this dimension as a dominant pillar of the study (Table 16).

Table 16. Dimension D one-sample t-test and Cohen’s d results.

Further analysis of Dimension D shows that the highest perceived significance is associated with D1, openness to learning new digital technologies, which shows a very large effect size (d = 1.42), identifying it as students’ strongest priority. Similarly high identification is observed for D4, motivation to develop AI competencies (d = 1.18), indicating that willingness to learn and internal motivation do not represent barriers to AI integration in education. Large effect sizes are also recorded for D5, creative thinking in AI contexts (d = 1.01), and D2, communication confidence in AI-related topics (d = 0.98), suggesting readiness not only to engage with AI but also to communicate about it and apply it creatively. Although D3, teamwork, and collaboration in technological and innovative tasks show the lowest effect size within the dimension (d = 0.90), it remains highly above average. Overall, Dimension D contains no deficient competencies and concentrates on the respondents’ strongest perceived capabilities.

Although all results for dimension E are found to be statistically significant, the effect sizes in this dimension are overall the lowest among all examined areas (Table 17).

Table 17. Dimension E one-sample t-test and Cohen’s d results.

Analysis of Dimension E shows that E3, perceived relevance of studies for future employability, is the most significant competency and the only one reaching a large effect size (d = 0.82), indicating that students view their studies as relevant despite rapid technological change. In contrast, E2, adaptability to labour market changes, records a moderate effect size (d = 0.54) and represents one of the weakest areas in the study, reflecting uncertainty about adapting to AI-driven labour market dynamics. E1, perceived readiness to work in AI-enabled environments (d = 0.74), and E4, awareness of AI-related learning and training opportunities (d = 0.64), exhibit moderate effect sizes. The findings reveal that students acknowledge baseline readiness; however, their confidence in practical transition and continuous professional development is markedly inferior to their theoretical understanding or intrinsic motivation.

5. Discussion

5.1. Summarisation

The viability of the conceptual framework (synthesised from existing AI literacy models, Miao et al., 2024; Jolly et al., 2024; SREB, 2025; Chee et al., 2024) is supported by the empirical evidence presented by this study. The internal consistency and the logical distribution of results across the five dimensions indicate that the framework is a robust instrument for measuring multidimensional AI competencies.

By synthesising the results across all five dimensions, a distinct pattern of the respondents’ AI competencies is observed: (i) a fundamental strength is identified in Dimension D (soft skills and attitudes towards AI) and Dimension A (understanding AI (AI literacy)), where high effect sizes (up to d = 1.42) demonstrate a high degree of openness, motivation, and ethical awareness. These findings suggest that a solid conceptual foundation is possessed by students who perceive the responsible development of AI as a priority; (ii) divergence is observed when moving from theoretical awareness to practical implementation. While conceptual principles (e.g., machine learning distinctions in Dimension B or data interpretation in Dimension C) are well-perceived, technical proficiency in active tasks—such as data visualisation, programming, or the active use of AI tools—exhibits noticeably lower effect sizes; (iii) the most critical area for systemic improvement is identified in Dimension E (labour market readiness). Despite their high motivation, the lowest levels of confidence are felt by students regarding their adaptability to AI-induced labour market changes and their awareness of further training opportunities. This creates a competency paradox: students are ethically aware and highly motivated, yet they feel technically underprepared and uncertain about their practical transition into an AI-augmented professional environment.

These results show that the soft and conceptual parts of AI literacy are well-defined, but it is necessary to work on improving hard technical skills and the ability to adapt to new situations in the workplace.

5.2. Policy Recommendations

Identifying weaknesses and strengths in students’ AI competencies establishes a foundation for subsequent policy recommendations. The context in which policy recommendations are formulated is based on the proposal to emphasise identified strengths and address weaknesses. Based on a sample of respondents, these recommendations are the foundation for curriculum creators of the Faculty of Operation and Economics of Transport and Communications of the University of Zilina. At the same time, it is important to note that this study and framework can be replicated on a sample of students from other faculties and universities worldwide. The same applies to policy recommendations stated in this section.

5.2.1. Dimension A—Understanding AI (AI Literacy)

To improve the comparatively diminished effect size in basic conceptual understanding (item A1, d = 0.74), curriculum developers ought to emphasise the reinforcement of formal theoretical frameworks. Instead of using definitions of AI that are based on intuition or the media, educational programmes need to make sure that a strong, standardised set of terms is set up at the beginning of the study. This conceptual anchoring is essential to transform students’ general awareness into a precise academic competency that serves as a prerequisite for more advanced technical discourse. The strength seen in ethical and social awareness (item A4, d = 1.25) gives teachers a chance to get students more involved in their lessons. Curriculum designers are urged to implement an ethics-driven instructional framework, wherein intricate technical concepts are presented through the perspective of social responsibility. High levels of confidence regarding the technological and economic context (items A2, d = 1.00, and A3, d = 1.10) should be utilised to foster deeper interdisciplinary integration. Since students demonstrate a clear understanding of the distinction between machine learning and classical programming, as well as AI’s role in the business environment, curricula should move toward complex case studies that synthesise these elements. These education modules should explicitly demonstrate how the specific algorithmic logic of machine learning directly influences economic value chains.

5.2.2. Dimension B—Technical and Digital Skills

Given the disparities in students’ technical preparedness, curriculum creators should focus on transforming passive data literacy into active technical creativity. The lowest effect size in the entire survey was for data visualisation and graph creation skills (item B2, d = 0.42). This means that it is important to provide specialised learning that focuses on how to communicate data visually. While students demonstrate proficiency in spreadsheet software (item B1, d = 1.03), they appear unable to transform this data into intelligible visual outputs. Consequently, a shift from fundamental spreadsheet processing to sophisticated visual storytelling tools is advised, facilitating students’ critical interpretation and presentation of results produced by AI systems. The need to strengthen direct practical experience with AI tools and the development of algorithmic thinking (items B3 and B4, d = 0.68 and 0.69) is identified. The results suggest that students suffer from a lack of practical implementation despite understanding the basic principles of machine learning (B5, d = 0.86). Curriculum designers should respond by incorporating laboratory exercises and project-based learning, enabling students to engage with AI not just theoretically, but by applying it to specific technical challenges. This strategy should also incorporate programming fundamentals in the context of automation, thus reconciling the disparity between theoretical comprehension and practical technical application, which students currently regard as a deficiency.

5.2.3. Dimension C—Cognitive and Analytical Abilities

Students exhibit proficiency in critically assessing digital technologies and their societal ramifications (item C4, d = 0.96). Curriculum designers ought to leverage this evaluative strength as a basis for fostering more rigorous analytical methodologies. The data show that students are good at socially auditing technology, but they need more structured help as they move to data-driven analytical problem-solving (C1, d = 0.68). Consequently, curricula must extend beyond discussions exclusively centred on ethics and commence the integration of applied analytics modules, in which students are obligated to substantiate their sociotechnical critiques with empirical evidence and systematic reasoning. The students are very sure they can understand data outputs (C3, d = 0.89), but they were not very good at making those same outputs (B2). Students are able to judge finished data, but they do not understand how it was made. To fix this, curriculum designers should hold workshops on reverse engineering. Students should be given finished AI-generated reports and asked to explain how the data sets and logical steps that went into them were put together during these sessions. This approach directly tackles the diminished effect size noted in Logical and Structured Decision-Making (C2, d = 0.79), necessitating that students engage with the complexities of the decision-making process rather than just the result.

5.2.4. Dimension D—Soft Skills and Attitudes Towards AI

With openness to learning new digital technologies (item D1, d = 1.42) and motivation for developing AI competencies (D4, d = 1.18) reaching high effect sizes, it is evident that students do not perceive AI as a threat to be avoided but as an opportunity to be embraced. This means that for people who make curricula, the traditional persuasion phase of educational change (getting students to see how important the technology is) is mostly not needed. Instead, the focus of educational strategy should be on high-engagement, proactive learning models that make use of this existing internal drive before it is lost due to boring or too theoretical lessons. A high level of communication confidence in AI-related topics (D2, d = 0.98) indicates that students are prepared to transition from passive learning to active discourse. Curriculum designers ought to integrate peer-to-peer teaching modules, necessitating that students articulate and justify their AI-related strategies. This builds on their confidence in communication to help them understand concepts better. Curriculum designers can turn individual confidence into a group academic culture that values technological fluency and public discussion of new ideas by creating an environment where students talk about and explain how AI works. Students also feel confident in creative thinking in the context of working with AI (D5, d = 1.01). Curricula should prioritise generative design and experimental prototyping rather than providing strictly guided instructions. Educators should offer open-ended challenges where AI tools are used as creative partners. This approach takes advantage of the fact that the students are ready to take action instead of just passively receiving information. By creating low-stakes environments where students can try out AI without worrying about failing, schools can turn this high level of motivation into real creative work, which will help the students feel more confident in their own strengths. Teamwork and collaboration in technological tasks (D3, d = 0.90) exhibits the relatively lowest value in this dimension; its high absolute score indicates a readiness for multidisciplinary cooperation. Curriculum designers should implement agile team projects that mimic real-world AI development environments. These projects should require students to combine their own creativity (D5) and communication skills (D2) with those of others. By treating Dimension D as a strategic dimension, institutions can use the students’ proactivity to pull them through the more demanding technical challenges identified in other dimensions.

5.2.5. Dimension E—Labour Market Readiness

The results from Dimension E identified a transition gap between academic perception and professional reality. While students maintain a high perceived relevance of studies for future employability (item E3, d = 0.82), this belief appears to be a general trust in the value of their degree rather than a specific confidence in their technical readiness. To address this, curriculum creators must move beyond traditional academic outcomes and incorporate detailed longitudinal AI-readiness frameworks. This involves mapping course-specific learning outcomes directly to emerging AI job roles, thereby transforming a general sense of relevance into concrete, career-oriented assurance. The most urgent area for systemic intervention is adaptability to labour market changes (E2, d = 0.54), which represents the second weakest link in the study. This moderate effect size suggests that students feel a profound sense of uncertainty regarding how AI will disrupt their specific career paths. Curriculum designers should respond by adding modules that teach not only AI tools but also meta-skills like prompt engineering for specialised workflows, understanding AI-augmented task automation, and dealing with changing employer expectations. By directly addressing these market dynamics in the curriculum, schools can help students feel less anxious about technology. A moderate level of perceived readiness to work in an AI-enabled environment (E1, d = 0.74) suggests that students need more immersive and professional experiences. It is recommended that curriculum designers establish cooperation with industry-led AI projects. These initiatives should simulate a professional environment where AI is a standard collaborator, moving the students’ sense of readiness from a baseline level to the high levels of certainty seen in their theoretical knowledge. This practical immersion is essential to bridge the gap between “knowing about AI” and “knowing how to work alongside AI.” Insufficient awareness of AI-related learning and training opportunities (E4, d = 0.64) highlights a need for better educational guidance. Since students do not have an overview of how to continue their development, curriculum creators should embed continuous learning roadmaps into the final stages of their studies. These roadmaps should give students curated access to micro-credentials, professional certifications, and specialised AI training platforms that are useful for their field. By institutionalising the awareness of lifelong learning, curriculum designers can ensure that students’ motivation (as seen in Dimension D) is directed toward the specific professional development opportunities required to remain competitive in an AI-augmented labour market.

5.3. Limitations of the Study and Agenda for Future Research

Despite the fact that the presented framework provides a comprehensive view of students’ perceived readiness, several limiting factors (and subsequent agenda for future research) must be considered when interpreting the results.

The primary limitation of the study is its reliance on self-assessment, which is inherently subjective. Self-assessment can be influenced by cognitive biases, such as social desirability or overconfidence, in areas that are popular in the media. For this reason, the results cannot be regarded as direct evidence of the respondents’ objective technical proficiency. A high level of perceived competence does not necessarily correlate with a high level of actual knowledge during in-depth testing. Conversely, a relatively low score in technical skills may reflect a higher degree of self-criticism among students who recognise the complexity of AI tools, rather than a total inability to use them. Consequently, the subsequent step involves enhancing the framework with more extensive research that analyses specific competencies in detail.
Another limitation is that the research is cross-sectional, meaning it only looks at students’ attitudes at one point in time, when technology was changing quickly. Future research should augment this integrated framework with longitudinal analysis to determine the relationship between subjective assessments of AI competencies and the actual employment outcomes of graduates in the labour market. In addition, the identified competencies applied within the competencies framework need to be validated within the foreseeable future so that it becomes clear how their structure is changing as AI advances rapidly.
The sample of respondents represents a further limitation of the study. The research involved first-year students from a single faculty with an economic-technical orientation at a selected university. Further research should broaden this scope by conducting measurements at faculties and universities with different academic focuses.
Literature review (Section 2) screening phase utilised AI (ChatGPT), which may not have identified every relevant study. The same limitation applies to subsequent phases related to the evaluation and inclusion of identified studies, which were carried out by the authors, where subjective judgement may have influenced the process. Future research should focus on revisiting the conducted literature review through the involvement of a larger number of researchers in the study evaluation process.
Policy recommendations (Section 5.2) represent a subjective construct of the authors of this study. Future research should focus on the empirical verification of the proposed recommendations. In addition, further research needs to focus on deeper policy development that goes beyond the scope of this study.

6. Conclusions

This study enhances the ongoing discourse on AI literacy by integrating and empirically demonstrating a multidimensional conceptual framework. The findings validate that the framework serves as a diagnostic tool, exhibiting the analytical precision necessary to distinguish among different levels of AI readiness. This research transcends a simplistic perspective of digital literacy by validating theoretical dimensions with empirical evidence, thus uncovering a multifaceted landscape of student competencies.

The findings illustrate that while the foundational pillars of AI literacy, specifically ethical awareness and intrinsic motivation, are firmly established, a significant divergence exists when moving toward practical and professional implementation. The identification of deficits in technical areas, such as data visualisation, and in strategic areas, such as labour market adaptability, indicates that current educational models are more successful in fostering informed observers than proficient practitioners. This gap suggests that high motivation and ethical alignment do not automatically translate into the procedural fluency required by an AI-augmented professional environment.

Consequently, this study underscores the necessity of a strategic shift in curriculum design within the faculty and university of interest. To effectively bridge the identified gaps, selected institutions must transition from purely conceptual instruction toward applied, project-based learning models that prioritise data-driven problem solving and technical tool proficiency. Moreover, the professional uncertainty felt by students necessitates the integration of AI-oriented career guidance to foster resilience in a volatile labour market.

Ultimately, the viability of the proposed framework lies in its practical applicability, as it provides curriculum creators with a precise map for targeted interventions. Even though the framework is based on a certain group of respondents, its empirical validation provides a basis for use and further testing in other student groups and institutional settings. By using students’ existing psychological readiness to fill in the gaps in their technical knowledge, schools can make sure that future professionals are not only morally grounded and motivated but also technically skilled enough to lead in the age of AI.

Author Contributions

Conceptualization: R.C. and N.S.; methodology: R.C.; software: R.C.; validation: R.C., R.N., N.S. and M.M.; investigation: R.C., R.N., N.S., M.M.; data curation: R.C.; writing—original draft preparation: R.C. and N.S.; writing—review and editing: R.C. and N.S.; visualization: R.C.; supervision: R.C.; project administration: N.S.; funding acquisition: N.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Slovak Research and Development Agency, grant number VV-MVP-24-0375.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki, and approved by The Ethics Board of the University of Zilina (protocol code EK 9/2025 and date of approval 18 November 2025).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.

Acknowledgments

During the preparation of this manuscript, the authors of this study also used AI chatbots, namely: (a) ChatGPT (model GPT-5.2) for purposes of the screening phase of literature review, (b) Gemini (model Gemini 3) for purposes of data analysis (descriptive statistics, one-sample t-tests, and Cohen’s d), where statistical processing was executed via scripts in Python 3.13 utilising the Pandas and SciPy libraries. The authors have reviewed and edited the output and take full responsibility for the content of this publication.

Conflicts of Interest

The authors declare no conflicts of interest.

Appendix A

Table A1. Query for identification of frameworks of AI-competencies. Asterisk (*) represents truncation operation. This operation will find any number of letters at the end of the word. For example student* will find students, student-based and so on. This type of operation was used to broaden the scope of search queries.

Table A2. Queries for identification of factors related to dimensions A–E.

Dimension A
WOS	Scopus
TS=(“AI literacy” OR “artificial intelligence literacy” OR (“artificial intelligence” AND literacy)) AND TS=(“labour market” OR “employer expectations” OR employability) AND TS=(“higher education” OR graduates)	TITLE-ABS-KEY((“AI literacy” OR “artificial intelligence literacy” OR (“artificial intelligence” AND literacy) OR (“artificial intelligence” AND understanding) OR (“artificial intelligence” AND knowledge)) AND (“labour market” OR “employer expectations” OR employability OR “professional practice”) AND (“higher education” OR students OR graduates))
Dimension B
WOS	Scopus
TS=(“digital skills” OR “technical skills” OR “data analysis skills” OR “programming skills” OR “AI skills”) AND TS=(“labour market” OR “employer expectations” OR employability) AND TS=(“higher education” OR graduates)	TITLE-ABS-KEY((“digital skills” OR “technical skills” OR “data analysis skills” OR “data visualisation” OR “programming skills” OR “coding skills”OR “AI skills” OR “machine learning basics” OR “use of AI tools”)AND(employability OR “labour market” OR “employer expectations” OR “professional practice”) AND (“higher education” OR students OR graduates))
Dimension C
WOS	Scopus
TS=(“cognitive skills” OR “analytical skills” OR “critical thinking” OR “problem solving skills”) AND TS=(employability OR “labour market” OR “employer expectations”) AND TS=(“higher education” OR students OR graduates)	TITLE-ABS-KEY((“cognitive skills” OR “analytical skills” OR “critical thinking” OR “problem solving skills” OR “data interpretation” OR “evidence-based decision making” OR “interpreting AI outputs”) AND (employability OR “labour market” OR “employer expectations” OR “professional practice”) AND (“higher education” OR students OR graduates))
Dimension D
WOS	Scopus
TS=(“soft skills” OR “non-cognitive skills” OR attitudes OR motivation OR creativity) AND TS=(“artificial intelligence” OR “digital technologies”) AND TS=(employability OR “labour market” OR “employer expectations”) AND TS=(“higher education” OR students OR graduates)	TITLE-ABS-KEY((“soft skills” OR “non-cognitive skills” OR attitudes OR motivation OR creativity OR adaptability OR teamwork OR collaboration OR “self-efficacy”) AND (“artificial intelligence” OR “digital technologies” OR “AI adoption”) AND (employability OR “labour market” OR “employer expectations” OR “professional practice”) AND (“higher education” OR students OR graduates))
Dimension E
WOS	Scopus
TS=(employability OR “labour market readiness” OR “workforce readiness”) AND TS=(“artificial intelligence” OR automation) AND TS=(adaptability OR reskilling OR upskilling OR “lifelong learning”) AND TS=(“higher education” OR students OR graduates)	TITLE-ABS-KEY((employability OR “labour market readiness” OR “workforce readiness” OR “job readiness”) AND (“artificial intelligence” OR automation OR “AI adoption”) AND (adaptability OR reskilling OR upskilling OR “lifelong learning” OR training) AND (“higher education” OR students OR graduates))

Appendix B

SLR in this study is methodologically composed of: (a) identifying studies related to the examined dimensions, taking into account open access scientific outputs indexed in the Web of Science (WOS) and Scopus databases; (b) subsequent screening and identification of studies that can be considered explicitly related to the examined dimension; (c) assessment of eligibility in terms of the relevance of the scientific outputs (by addressing the question: Does the examined study describe factors influencing the examined dimensions?); and (d) extraction of information and inclusion of studies that provided a sufficient level of answers to the question outlined in point (c) in the final list of included studies.

The total number of analysed studies is presented in Table 3. Screening (step (b)) was conducted using the ChatGPT (version GPT 5.2) application as follows: based on research queries related to each dimension and database (Appendix A, Table A2), a database of records was generated and exported in the form of a .csv file. The file was uploaded into the ChatGPT application, where, using structured commands, records whose titles and abstracts explicitly addressed the examined dimensions A–E were identified. Subsequently, steps (c) and (d) were carried out under human supervision, in which the authors themselves evaluated the selected studies and made the final selection for further processing.

Table A3. Systematic literature review.

Appendix C

Table A4. Competency labels for survey items and questions in the questionnaire.

Survey Item	Competency Label	Question in Questionnaire
A1	Basic understanding of artificial intelligence (AI literacy—conceptual understanding)	9. I understand what AI is. I know what the term AI means and what tasks it can solve.
A2	Applied understanding of AI use in the economy and business	10. I know examples of the use of AI in the economy or business. I can list situations where companies use AI.
A3	Understanding the distinction between machine learning and classical programming	11. I can explain the difference between machine learning and classical programming. I understand that machine learning learns from data, while classical programming uses rules.
A4	Ethical and social awareness of artificial intelligence	12. I understand the social and ethical issues associated with AI. I know what risks or dilemmas AI can bring (e.g., privacy).
B1	Basic digital data skills and spreadsheet proficiency	13. I can work with spreadsheets and data analysis tools (e.g., Excel, SPSS). I know the basic functions and can process data.
B2	Data visualisation and graph creation skills	14. I have experience with creating graphs and data visualizations. I have already created graphs or visual representations of data.
B3	Basic programming and algorithmic thinking skills	15. I can write simple computer code or understand the logic of algorithms. I understand the basic principles of programming.
B4	Practical experience with AI tools	16. I have used AI tools (e.g., ChatGPT, chatbots, systems for translation and editing text—Grammarly, DeepL and the like). I have practical experience using AI applications.
B5	Basic understanding of machine learning principles	17. I understand the basics of how machine learning algorithms work, which are the essence of AI. I know that algorithms learn from data and adapt.
C1	Data-driven analytical problem-solving skills	18. I can solve complex problems by analyzing data and evidence. I can use data to propose solutions to a problem.
C2	Logical and structured decision-making	19. I use logical and structured thinking in decision-making. I make decisions systematically and based on facts.
C3	Interpretation of data outputs, graphs, and statistics	20. I can interpret graphs, statistics or AI outputs. I understand what graphs or statistics mean.
C4	Critical evaluation of digital technologies and their societal impact	21. I think critically about digital technologies and their impact on society. I can evaluate the benefits and risks of digital technologies.
D1	Openness to learning new digital technologies	22. I am open to learning new digital tools. I have no problem learning new technologies.
D2	Communication confidence in AI-related topics	23. I feel confident explaining AI topics to other people. I am confident in communicating and explaining the basics of working with AI.
D3	Teamwork and collaboration in technological and innovative tasks	24. I enjoy working in a team on technological or innovative tasks. I enjoy collaborating on technical or creative projects.
D4	Motivation for developing AI and digital competencies	25. I am motivated to improve my skills in AI and digital technologies. I want to develop myself further in this area.
D5	Creative thinking in the context of working with AI	26. I believe that creativity is important when working with AI. I think that ideas and creativity are important in AI.
E1	Perceived readiness to work in an AI-enabled environment	27. I think I have sufficient skills to work in an environment where AI is used. I believe I would be able to handle tasks related to AI.
E2	Adaptability to labour market changes driven by AI	28. I feel ready to adapt to changes in the job market due to AI. I am ready to learn new things as technology develops.
E3	Perceived relevance of studies for future employability	29. My studies, whether on my own or at school, help me develop the skills needed for the future job market. My own initiative to learn and the opportunities at school give me skills that I will use in practice.
E4	Awareness of AI-related learning and training opportunities	30. I know where to find educational opportunities in the field of AI (e.g., online courses, training). I know the platforms where you can learn to work with AI.

Appendix D

Table A5. Questionnaire—Part 1. Demographics.

1. Gender	A. Male B. Female C. Other D. Prefer not to disclose
2. Age	Please state your age (open question)
3. What type of secondary school did you attend? (Select the option that best corresponds to your studies.)	A. Grammar School/Gymnasium (general, languages, bilingual, sports) B. Business Academy/Secondary Business School (economics, accounting, entrepreneurship) C. Secondary Technical School—Engineering/IT (electrotechnics, informatics, mechatronics, mechanical engineering, digital media graphics) D. Vocational Secondary School—Economics-related (excluding Business Academy) (economic lyceum, hotel academy, tourism, services management) E. Vocational Secondary School—Humanities or Social Sciences (pedagogical, social work, mass media, legal/administrative) F. Vocational Secondary School—Other fields (gastronomy, healthcare, agriculture, crafts, services) G. Conservatory (music, dance, singing, acting, musical-dramatic arts) H. Other (please specify):
4. In which region did you attend secondary school?	A. Bratislava Region B. Trnava Region C. Trenčín Region D. Nitra Region E. Žilina Region F. Banská Bystrica Region G. Prešov Region H. Košice Region I. Outside the Slovak Republic (please specify):
5. Are you looking to pursue a career in a field that utilizes artificial intelligence?	1. Strongly disagree 2. Disagree 3. Neither agree nor disagree 4. Agree 5. Strongly agree
6. To what degree did you utilize digital technologies, including computers, software, and online tools, throughout your secondary education?	1. Never 2. Rarely 3. Sometimes 4. Often 5. Almost always
7. To what degree did your secondary school provide access to computer science or digital technology courses?	1. We had no such subjects 2. We had them only rarely 3. We had them occasionally (e.g., for 1–2 years) 4. We had them regularly across multiple years 5. We had them every year
8. How would you rate your current English language proficiency (according to levels A2–C2)?	1. A2—Elementary/Waystage 2. B1—Intermediate 3. B2—Upper-Intermediate 4. C1—Advanced 5. C2—Proficiency

Table A6. Questionnaire—Part 1. Demographics.

9. I understand what AI is. I know what the term AI means and what tasks it can solve.	1	2	3	4	5
10. I know examples of the use of AI in the economy or business. I can list situations where companies use AI.	1	2	3	4	5
11. I can explain the difference between machine learning and classical programming. I understand that machine learning learns from data, while classical programming uses rules.	1	2	3	4	5
12. I understand the social and ethical issues associated with AI. I know what risks or dilemmas AI can bring (e.g., privacy).	1	2	3	4	5
13. I can work with spreadsheets and data analysis tools (e.g., Excel, SPSS). I know the basic functions and can process data.	1	2	3	4	5
14. I have experience with creating graphs and data visualizations. I have already created graphs or visual representations of data.	1	2	3	4	5
15. I can write simple computer code or understand the logic of algorithms. I understand the basic principles of programming.	1	2	3	4	5
16. I have used AI tools (e.g., ChatGPT, chatbots, systems for translation and editing text—Grammarly, DeepL and the like). I have practical experience using AI applications.	1	2	3	4	5
17. I understand the basics of how machine learning algorithms work, which are the essence of AI. I know that algorithms learn from data and adapt.	1	2	3	4	5
18. I can solve complex problems by analyzing data and evidence. I can use data to propose solutions to a problem.	1	2	3	4	5
19. I use logical and structured thinking in decision-making. I make decisions systematically and based on facts.	1	2	3	4	5
20. I can interpret graphs, statistics or AI outputs. I understand what graphs or statistics mean.	1	2	3	4	5
21. I think critically about digital technologies and their impact on society. I can evaluate the benefits and risks of digital technologies.	1	2	3	4	5
22. I am open to learning new digital tools. I have no problem learning new technologies.	1	2	3	4	5
23. I feel confident explaining AI topics to other people. I am confident in communicating and explaining the basics of working with AI.	1	2	3	4	5
24. I enjoy working in a team on technological or innovative tasks. I enjoy collaborating on technical or creative projects.	1	2	3	4	5
25. I am motivated to improve my skills in AI and digital technologies. I want to develop myself further in this area.	1	2	3	4	5
26. I believe that creativity is important when working with AI. I think that ideas and creativity are important in AI.	1	2	3	4	5
27. I think I have sufficient skills to work in an environment where AI is used. I believe I would be able to handle tasks related to AI.	1	2	3	4	5
28. I feel ready to adapt to changes in the job market due to AI. I am ready to learn new things as technology develops.	1	2	3	4	5
29. My studies, whether on my own or at school, help me develop the skills needed for the future job market. My own initiative to learn and the opportunities at school give me skills that I will use in practice.	1	2	3	4	5
30. I know where to find educational opportunities in the field of AI (e.g., online courses, training). I know the platforms where you can learn to work with AI.	1	2	3	4	5

References

Alam, M. J., Noman, S., Mujib, M. N. I., & Khan, W. S. (2025). An assessment of graduates’ skills gap for sustainable employability during the 4IR in Bangladesh. Social Sciences & Humanities Open, 12, 101780. [Google Scholar] [CrossRef] [Scilit]
Aljaaidi, K. S., Abdulmajeed, I. A. I., Bafaqeer, S. M., & Habeeb, S. (2025). Mapping SDGs’ 4 and 8 through enhancing technological skills for students’ employability. International Journal of Data and Network Science, 9(1), 57–76. [Google Scholar] [CrossRef] [Scilit]
Aly, A., Elazazy, A., & Sharaf, N. A. H. (2025, February 19–21). Integrating generative AI in architectural education: A comparative study of traditional, stock LLMs, and custom tools [Conference Contribution]. International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (Vol. 1, pp. 415–420), Lisbon, Portugal. [Google Scholar]
Al Zaabi, A., Al Maskari, S., & Aal Abdulsalam, A. (2023). Are physicians and medical students ready for artificial intelligence applications in healthcare? Digital Health, 9, 20552076231152167. [Google Scholar] [CrossRef] [Scilit]
Bartolomé, J., del Río, I., Martínez, A., Aranguren, A., Laña, I., & Alloza, S. (2025). Game on: Exploring the potential for soft skill development through video games. Information, 16(10), 918. [Google Scholar] [CrossRef] [Scilit]
Başal, M., Süzen, E., Şarkbay, Ö. F., Öztürk, E., & Keke, M. E. (2025). Sustainable supply chain training: Digital tools for transportation modules. E3S Web of Conferences, 646, 00041. [Google Scholar] [CrossRef] [Scilit]
Bedoya-Guerrero, A., Basantes-Andrade, A., Rosales, F. O., Naranjo-Toro, M., & León-Carlosama, R. (2024). Soft skills and employability of online graduate students. Education Sciences, 14(8), 864. [Google Scholar] [CrossRef] [Scilit]
Bhatti, M. A., Mat Saat, S. A., Aleidan, M. M., Al Murshidi, G. H. M., Alyahya, M., & Juhari, A. S. (2022). Are business graduates’ employability skills and learning/teaching techniques universal? Exploring the role of culture. Sustainability, 14(5), 3069. [Google Scholar] [CrossRef] [Scilit]
Bondin, A., & Zammit, J. P. (2025). Education 4.0 for Industry 4.0: A mixed reality framework for workforce readiness in manufacturing. Multimodal Technologies and Interaction, 9(5), 43. [Google Scholar] [CrossRef] [Scilit]
Botha, Z., & Dreyer, L. M. (2022). Digital technology as an alternative approach to teaching skills to learners with severe intellectual disabilities. Tydskrif vir Geesteswetenskappe, 62(2), 364–391. [Google Scholar] [CrossRef] [Scilit]
Camacho-Zuñiga, C., Salas-Maxemín, S., Valle-Arce, A. P., Caratozzolo, P., & Chans, G. M. (2025). Toward a continuous learning educational model: Insights from the experience of a Mexican private university. Frontiers in Education, 10, 1485034. [Google Scholar] [CrossRef] [Scilit]
Chee, H., Ahn, S., & Lee, J. (2024). A competency framework for AI literacy: Variations by different learner groups and an implied learning pathway. British Journal of Educational Technology, 56(5), 2146–2182. [Google Scholar] [CrossRef] [Scilit]
Constantin, A., & Iacob, M. (2025). Higher education in the age of AI: ChatGPT’s impact on students’ academic activities. Education and Information Technologies, 19, 3360–3376. [Google Scholar] [CrossRef] [Scilit]
Corrales-Herrero, H., & Rodríguez-Prado, B. (2024). Mapping the occupations of recent graduates: The role of academic background in the digital era. Research in Higher Education, 65, 1853–1882. [Google Scholar] [CrossRef] [Scilit]
Creely, E., Barnes, M., Tour, E., Henderson, M., Waterhouse, P., Peña, M. A., & Patel, S. V. (2025). Exploring attitudes to generative AI in education for English as an additional language (EAL) adult learners. ReCALL, 37(2), 174–190. [Google Scholar] [CrossRef] [Scilit]
Dumitru, D., & Halpern, D. F. (2023). Critical thinking: Creating job-proof skills for the future of work. Journal of Intelligence, 11(10), 194. [Google Scholar] [CrossRef] [Scilit]
El Hosayny, O., Razkane, H., & Yeou, M. (2025). Extended TAM-based acceptance of artificial intelligence for supporting critical thinking in academic research. Humanities and Social Sciences Communications, 12, 176. [Google Scholar]
Ferro, E., Martínez-Cantos, J. L., & Gómez-Hernández, J. A. (2023). Gender gap in professional digital competence: Construction and initial validation of an instrument for its measurement. Feminismos, 45, 137–162. [Google Scholar]
Fleming, G. C., Klopfer, M., Katz, A., & Knight, D. (2024). What engineering employers want: An analysis of technical and professional skills in engineering job advertisements. Journal of Engineering Education, 113(2), 251–279. [Google Scholar] [CrossRef] [Scilit]
Göltl, K., Ambros, R., Dolezal, D., & Motschnig, R. (2024). Pre-service teachers’ perceptions of their digital competencies and ways to acquire those through their studies and self-organized learning. Education Sciences, 14(9), 951. [Google Scholar] [CrossRef] [Scilit]
Grama, A., & Todericiu, R. (2025). The evolution of skill dynamics in the context of the future of work. Systems, 13(1), 6. [Google Scholar] [CrossRef] [Scilit]
He, X. (2025). A novel MCDM method for assessing student employability psychology using complex Pythagorean hesitant fuzzy Sugeno–Weber aggregation. Scientific Reports, 15(1), 26676. [Google Scholar] [CrossRef] [Scilit]
Hrnjić, A., Pilav-Velić, A., Hadžiahmetović-Milišić, A., & Dedović, L. (2025). The future of work in transition economies: Integrating AI, digital skills, and employability in higher education. South East European Journal of Economics and Business, 20(2), 85–99. [Google Scholar] [CrossRef] [Scilit]
Islam, M. A. (2022). Industry 4.0: Skill set for employability. Social Sciences & Humanities Open, 6, 100280. [Google Scholar] [CrossRef] [Scilit]
Jolly, S., Potter, T., Cosby, M., Meyr, N., Balakrishnan, V., & Jayaram, G. G. M. (2024). LEADERS: Skilling/reskilling framework for the future AI workforce—Demonstrated by a healthcare case study (Microsoft 2023–24 ExecMAP Final Report). University of Michigan—Ross School of Business. [Google Scholar]
José-García, A., Sneyd, A., Melro, A., Ollagnier, A., Tarling, G., Zhang, H., Stevenson, M., Everson, R., & Arthur, R. (2023). C3-IoC: A career guidance system for assessing student skills using machine learning and network visualisation. International Journal of Artificial Intelligence in Education, 33(4), 1094–1121. [Google Scholar] [CrossRef] [Scilit]
Kamaruzaman, F. M., Nik Othman, N. N. J., Omar, M., & Mohd Zaid, M. F. A. (2025). Future generic skills for technical vocational education graduates. International Journal of Learning, Teaching and Educational Research, 24(3), 312–332. [Google Scholar] [CrossRef] [Scilit]
Liu, Y., Hua, H., & Zhou, N. (2025, July 11–13). XGBoost regression of employment competencies and satisfaction among University B’s 2025 liberal arts graduates [Conference Contribution]. 2025 3rd International Conference on Educational Knowledge and Informatization (EKI 2025) (pp. 1–5), Chongqing, China. [Google Scholar]
Mahazan, M. H., & Ismail, H. H. (2025). Artificial intelligence tools as catalysts of improved spoken English: A systematic review of current applications and challenges. International Journal of Learning, Teaching and Educational Research, 24(12), 675–701. [Google Scholar] [CrossRef] [Scilit]
Mahdin, H., Nurwarsito, H., Baharum, Z., Kamri, K. A., Hassan, A., Haw, S.-C., & Arshad, M. S. (2025). Predictive analytics for employability in Malaysian TVET with a hybrid of regression and clustering methods. International Journal on Informatics Visualization, 9(5), 1816–1821. [Google Scholar] [CrossRef] [Scilit]
Malikaa, O., & Badra, E. (2025). The internationalization of higher education and labor market transformations: Study of current and future digital capabilities. Multidisciplinary Science Journal, 7, e2025088. [Google Scholar] [CrossRef] [Scilit]
Matiwane, B. Y., & Olaitan, O. (2025). Critical success factors for integrating AI tools into university curricula for workforce readiness. International Journal of Learning, Teaching and Educational Research, 24(8), 91–111. [Google Scholar] [CrossRef] [Scilit]
Miao, F., Shiohira, K., & Lao, N. (2024). AI competency framework for students. UNESCO. [Google Scholar]
Morgan, N., Yaqoob, M., & Jones, M. A. (2025). Using digital escape rooms to enhance data analysis skills and student experience in a higher education human physiology module. Current Research in Physiology, 8, 100165. [Google Scholar] [CrossRef] [Scilit]
OECD. (2023). OECD employment outlook 2023: Artificial intelligence and the labour market. OECD Publishing. [Google Scholar]
Opesemowo, O. A. G., Iwintolu, R. O., Odeyemi, T. S., Afolabi, A. O., & Afolabi, O. J. (2025). Exploration of employability skills among undergraduate students in the era of the fourth industrial revolution. Humanities and Social Sciences Communications, 12, 767. [Google Scholar]
Porjazoska Kujundziski, A., Domazet, E., Kamberaj, H., Rahmani, D., Abazi Feta, A., Lopez Valverde, F., Gálvez Rojas, S., Petlenkov, E., Vassiljeva, K., Štajduhar, I., & Hagen, T. (2024). Transversal skills in applied artificial intelligence—The case of the financial sector. E-Mentor, 2(104), 82–90. [Google Scholar] [CrossRef] [Scilit]
Portocarrero Ramos, H. C., Cruz Caro, O., Sánchez Bardales, E., Quiñones Huatangari, L., Campos Trigoso, J. A., Maicelo Guevara, J. L., & Chávez Santos, R. (2025). Artificial intelligence skills and their impact on the employability of university graduates. Frontiers in Artificial Intelligence, 8, 1629320. [Google Scholar] [CrossRef] [Scilit]
Pronello, C. (2026). Artificial intelligence and its impact on the current and future transport workforce. Transportation Research Interdisciplinary Perspectives, 36, 101802. [Google Scholar] [CrossRef] [Scilit]
Rampersad, G. C. (2020). Robot will take your job: Innovation for an era of artificial intelligence. Journal of Business Research, 116, 68–74. [Google Scholar] [CrossRef] [Scilit]
Ravšelj, D., Keržič, D., Tomaževič, N., Umek, L., Brezovar, N., Iahad, N. A., Abdulla, A. A., Akopyan, A., Segura, M. W. A., AlHumaid, J., & Allam, M. F. (2025). Higher education students’ perceptions of ChatGPT: A global study of early reactions. PLoS ONE, 20(2), e0315011. [Google Scholar] [CrossRef] [Scilit]
Reich, Y. (2025). Learning to use generative AI and using it to improve learning: A systems engineering research seminar case study. Systems, 13(11), 1006. [Google Scholar] [CrossRef] [Scilit]
Roberts, J. C. (2025). Data meets creativity: Authentic learning through data art design and exhibition. Computers & Graphics, 129, 104248. [Google Scholar] [CrossRef] [Scilit]
Schmidt, D. A., Alboloushi, B., Thomas, A., & Maguire, P. (2025). Integrating artificial intelligence in higher education: Perceptions, challenges, and strategies for academic innovation. Computers and Education: Artificial Intelligence, 6, 100274. [Google Scholar] [CrossRef] [Scilit]
Silva, J. B., Silva, I. N., & Bilessimo, S. (2020). Technological structure for technology integration in the classroom, inspired by the maker culture. Journal of Information Technology Education: Research, 19, 167–204. [Google Scholar] [CrossRef] [Scilit]
Singh, A., Singh, H. P., & Alam, M. F. (2026). The impact of AI tools enabled student engagement on academic performance, employability, and empowerment in Ha’il Region, Saudi Arabia. Acta Psychologica, 262, 106103. [Google Scholar] [CrossRef] [Scilit]
Southern Regional Education Board. (2025). Skills for an AI-ready workforce. SREB. [Google Scholar]
Tolegen, M. A., Baymuhambetova, B., Rovnyakova, I., Radchenko, N., Sakhariyeva, S., & Anafia, P. (2025). Crowdsourcing in Kazakhstan’s higher education in the system of dual education as predictor of universal compe-tencies. International Journal of Evaluation and Research in Education, 14(6), 4614–4627. [Google Scholar] [CrossRef] [Scilit]
Toma, S. G., & Hudea, O. S. (2024). Generation Z students’ perceptions on the abilities, skills and competencies required in the age of artificial intelligence systems. Amfiteatru Economic, 26(65), 162–180. [Google Scholar] [CrossRef] [Scilit]
Tsiligiris, V., & Bowyer, D. (2021). Exploring the impact of 4IR on skills and personal qualities for future accountants. Accounting Education, 30(6), 621–649. [Google Scholar] [CrossRef] [Scilit]
Velu, S. R. (2023). Design thinking approach for increasing innovative action in universities: ICT’s mediating effect. Sustainability, 15(1), 24. [Google Scholar] [CrossRef] [Scilit]
Vrbančič, F., & Kocijančič, S. (2024). Strategy for learning microcontroller programming—A graphical or a textual start? Education and Information Technologies, 29, 5115–5137. [Google Scholar] [CrossRef] [Scilit]
Waring, P. (2024). Artificial intelligence and graduate employability: What should we teach Generation AI? Journal of Applied Learning and Teaching, 7(1), 22–25. [Google Scholar]
Wong, G. S., & Abdullah, N. S. (2025). Job readiness level to address 4.0 skills: An empirical study on TVET students in electrical and electronics field. Journal of Technical Education and Training, 17(1), 62–76. [Google Scholar] [CrossRef] [Scilit]
Xu, L., Zhang, J., Ding, Y., Sun, G., Zhang, W., Philbin, S. P., & Guo, B. H. (2022). Assessing the impact of digital education and the role of the big data analytics course to enhance the skills and employability of engineering students. Frontiers in Psychology, 13, 974574. [Google Scholar] [CrossRef] [Scilit]
Yurdunkulu, A., Bulut, M. A., & Göçen, A. (2025). From academics to Aidemics: Unpacking the human–AI collaboration in higher education. Societies, 15(1), 6. [Google Scholar]
Zhan, S., & Wang, L. (2024). Integrating generative AI in higher education: Challenges, opportunities, and innovations in assessment practices within electrical engineering. Education Research and Perspectives, 51, 55–81. [Google Scholar] [CrossRef] [Scilit]